Browse Prior Art Database

Method for measuring run-to-run control system stability and robustness performance for the photolithographic process

IP.com Disclosure Number: IPCOM000130481D
Publication Date: 2005-Oct-25
Document File: 4 page(s) / 95K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for measuring run-to-run control system stability and robustness performance for the photolithographic process. Benefits include improved functionality and improved performance.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 39% of the total text.

Method for measuring run-to-run control system stability and robustness performance for the photolithographic process

Disclosed is a method for measuring run-to-run control system stability and robustness performance for the photolithographic process. Benefits include improved functionality and improved performance.

Background

              Analyzing and quantitatively measuring the performance of the photolithographic process in the presence of model mismatch and unknown disturbances is difficult. Conventionally, stability analysis is performed with a fixed statistical disturbance model, such as the integral moving average. For disturbances without information about its statistical model and probability distribution, no conventional solution exists.

              Two measures of control system performance are the stability tolerance to the model mismatch and the impact of unknown noise to the output performance. The stability tolerance to the model mismatch is a measure of the stability tolerance when the estimated process model does not match the actual process model. This situation is typical because the actual process model is generally unknown. However, run-to-run controllers are derived based on estimated process models. When an estimated model is much different from the actual model, a controller could become incorrect for the real system and an unstable process can occur.

              Conventional analysis of a controlled semiconductor manufacturing process is not able to quantitatively measure the effect of unknown disturbance with a large range of different frequencies. Information about statistical model or probability distribution is assumed for noise and disturbances.

              Critical dimension (CD), a measurement of lines and ditches in the circuit, is decisive to the speed of semiconductor products. Stable and uniform CD output in a manufacturing process is critical to reduce defect rates and production cost. In a photolithographic process, the CD output is impacted by a variety of factors. Some of them are known and some others are unknown due to little or no information. The dose and focus of the process are adjusted based on the output of the previous runs. This process model can never be 100% accurate. The model mismatch causes control variation, which results in CD variation in the product.

General description

              The disclosed method includes quantitative analysis and measurement of the performance of a run-to-run control photolithography process in the presence of model mismatch and unknown disturbances. The method derives the range of the model mismatch in which a photolithographic process is able to maintain its stability. The range of stability tolerance is an indicator that enables engineers to select a controller capable of handling increased model mismatch.

              The impact of unknown noise on the output performance indicates how sensitive the output can be to the unknown noise in the process. In the disclos...